Learn the Craft of Getting Funded

4.8K followers + Follow
More artists are getting funded with Grantcraft.art ★ New Grants added weekly ★ Join The Grantcraft Brief ★

Outsourcing Your Brain? Think for Yourself in the AI Age

Poem Johnson Avatar
Poem Johnson

0views

Are you outsourcing your brain to artificial intelligence? Generative AI tools are redefining cognitive offloading – the use of external aids (writing, calculators, GPS) to perform mental tasks. Unlike calculators or search engines, large language models generate novel content and structure ideas, offloading not just retrieval but reasoning itself. Emerging research (2024–26) suggests AI reliance can reduce mental effort and memory encoding.
For example, MIT EEG studies show ChatGPT users have significantly lower brain engagement: up to 55% weaker neural connectivity than unaided writers. Generative-AI “copy-paste” use correlates with poorer recall (Kosmyna et al., 2025: 83% of AI-assisted writers could not remember text they just wrote) and lower critical-thinking scores. Secondary evidence from surveys likewise links heavy AI tool use to lower critical thinking (e.g. Gerlich 2025, r≈–0.5).

However, the picture is nuanced. Tools have always changed cognition (writing, calculators, search) and can amplify abilities when used wisely. Recent experiments show structured AI use can boost outcomes: one field trial found ChatGPT made employees more creative – but only those who actively planned and monitored their prompts. The key distinction is often how AI is used, not that it exists. When learners outline their own ideas and critically review AI outputs, AI can scaffold higher-order thinking; conversely, unguided dependence breeds passivity.

This report reviews contemporary (2024–26) evidence. We define cognitive offloading in context (Section 1), contrast AI with prior tools (2), and summarize experiments on memory, attention, problem-solving, creativity and metacognition (3–4). We examine neuroscience findings on neural engagement (5) and situate AI dependence in historical perspective (6). Social factors – algorithmic authority, recommendation algorithms, “learned helplessness” – are also reviewed (7). Across topics we grade evidence (Strong/Moderate/Preliminary). We conclude with open questions, implications for education/art/democracy, and recommendations to preserve cognitive autonomy (8).

1. Cognitive Offloading: Concept and Evolution

  • Definition: Cognitive offloading means delegating tasks to external tools to lighten mental workload. It has long roots: writing systems (memory offload), abacuses/calculators (math), maps/GPS (navigation). Iconic examples: Socrates (via Plato) warned writing would erode memory; modern “Google effect” shows people recall facts less when they know answers are online. GPS users likewise can weaken spatial memory.
  • Extended Mind Theory: Theorists (Clark & Chalmers 1998 et al.) argue cognition often spans brain and world. Clark (2025) stresses that humans have always created “new ways of thinking” by integrating tools: e.g. “we have been building new ways of thinking… [with] embodied action to make the most of our worlds.” Writing, calculators, and even finger-counting “become part of how we think.” In the AI era, Clark notes our minds form “delicately interwoven wholes” of brain, body and digital tools. Crucially, Clark emphasizes skill in offloading: “we need to become experts at deciding the extent to rely on various forms of ‘cognitive offloading’.”
  • Contemporary framing: Frontiers (Jose et al. 2025) echoes this: “Cognitive offloading refers to the utilization of external aids to achieve cognitive tasks (Risko & Gilbert, 2016)… AI-powered tools make it easier, but they also reduce opportunities for active recall and problem-solving.” Offloading is not inherently bad – it trades effort for efficiency – but excessive dependence can diminish engagement. This report uses “cognitive offloading” in the modern sense of shifting tasks (recall, calculation, even idea-generation) to AI tools.

2. AI vs. Prior Cognitive Technologies

Generative AI tools qualitatively differ from earlier tech: they create new content on demand and do not merely retrieve stored answers. As one analysis notes: “Unlike a calculator… or a search engine… large language models can generate novel text, structure arguments, synthesize information, and mimic reasoning.” Key distinctions:

  • Generative vs. Procedural: Calculators and spell-checkers do bounded, well-defined operations. In contrast, LLMs hallucinate (“built to sound plausible, not necessarily correct”) and can produce arbitrary textual output. Santana (2025) warns “calculators don’t hallucinate… [LLMs] are built to sound plausible, not to be right.”
  • Scope of Delegation: Calculators offload arithmetic; GPS offloads navigation; search offloads information lookup. LLMs, however, have no fixed domain. In principle any cognitive task (essay writing, coding, problem-solving) can be handed off, which blurs where human effort ends. “Calculators offload clearly defined tasks. With LLMs, delegation is murkier… Where does cognitive effort end and offloading begin?” Accountability and authorship become opaque.
  • Interface and Trust: Generative AI’s conversational interfaces and anthropomorphic style foster undue authority. Whereas we treat a calculator’s output skeptically, AI outputs often look like advice from an “expert”. Hardcastle (2025) warns of “algorithmic authority”: students risk outsourcing judgment itself to Big Tech’s models. Unlike past tools, ChatGPT-style agents may shift users’ epistemic framework – they decide what counts as a good answer. This feeds a subtle but pervasive trust in AI, furthering offloading beyond simple recall.

3. Cognitive Effects of AI Reliance

Memory & Recall: Multiple studies indicate heavy AI use can weaken memory encoding. In the MIT Media Lab EEG experiment (Kosmyna et al. 2025), 83% of AI-assisted writers could not recall a phrase they had just written with ChatGPT. Frontiers (Jose et al. 2025) cite Bai et al. (2023): “While AI enhances personalized learning, excessive reliance may reduce cognitive engagement and long-term retention.” In one classroom test, students who used AI extensively showed memory decline: after initial “pretesting” aids, “prolonged AI exposure led to memory decline.” These findings (both lab and educational surveys) tentatively point to an inverse relationship between AI offloading and memory consolidation (Evidence Strength: Moderate).

Cognitive Effort and Engagement: When AI handles core tasks, people expend less mental effort. The MIT EEG study found ChatGPT users wrote 60% faster but with 32% less “cognitive load” than unaided writers. Neuroscientist Kosmyna et al. report LLM users had weaker brain connectivity in task-related regions, while brain-only writers showed more distributed engagement. A science-communication piece summarized: “ChatGPT users had the lowest brain engagement and consistently underperformed at neural… levels” In sum, generative AI tends to lighten momentary effort (we think less while working) but at the cost of diminished neural activation (Evidence Strength: Strong given EEG data).

Critical Thinking & Problem Solving: Studies suggest AI reliance can erode critical scrutiny. Ododo et al. (2024, cited in) found vocational students using AI often accepted information passively, “without critical scrutiny”. In surveys, more frequent AI use correlates with lower critical-thinking scores. For instance, Gerlich’s large survey (N=666) found a significant negative correlation between AI tool use and critical thinking, mediated by offloading. Experimental findings align: participants who used ChatGPT mindlessly often imported flawed logic or failed to critically evaluate AI’s suggestions. These patterns imply generative AI can bypass active reasoning, making users reliant on the tool’s outputs, especially when left unguided. (Evidence Strength: Moderate. Many studies are correlational or limited-in-time; causality is plausible but hard to prove.)

Decision-Making & Metacognition: There is scant direct research on decision-making, but theoretical concerns abound. Automated decision aids (like AI copilots) may induce “automation complacency” and under-trusting self-judgment. One analogy: pilots overly dependent on autopilot can lose situational awareness. Early surveys (e.g. Tian & Zhang 2025) indicate frequent AI use associates with cognitive fatigue and diminished judgment, but rigorous RCTs are lacking. Importantly, structured AI use (see below) can mitigate these risks.

Attention & Focus: The multitasking, prompt-focused nature of AI use may fragment attention. Though not yet extensively studied, experts warn that on-demand AI can promote superficial reading and disrupted concentration – an extreme form of the existing “attention economy”. Some preliminary surveys hint that heavy AI users report more distraction and reduced ability to sustain effort on complex tasks (Sources: workplace surveys, anecdotal). More empirical work is needed (Evidence: Preliminary).

4. Educational Outcomes and Independence

Generative AI in education has two facets: learning aids and outsourcing crutches. Recent studies highlight this duality:

  • Learning Performance: A large quasi-experiment (270 students, economics course) found ChatGPT “users completed tasks 30% faster than non-users”, but with no improvement in solution quality. In fact, although AI users passed slightly more exams (higher throughput), their grades were not higher. The authors conclude that AI enabled greater efficiency but no deeper learning gain: students “worked more efficiently but did not achieve higher-quality responses.” This suggests a trade-off: AI can boost productivity (Strong Evidence).
  • Skill Development: Evidence on foundational skills is mixed. In one field experiment (Lu et al. 2025), 250 employees using ChatGPT showed creativity gains, but only if they had strong planning and self-monitoring (metacognitive) skills. Those lacking such strategies saw no benefit. Similarly, educational case studies find that when students use AI as a tool with reflection, it can scaffold learning (e.g. prompting students to outline answers before AI help improves engagement). By contrast, unguided use often leads to rote copying. A Harvard interview quotes an education expert: “No learning occurs unless the brain is actively engaged… [with AI] that is not going to occur if you just ask ChatGPT ‘give me the answer’.”
  • Metacognition: AI use can degrade students’ ability to self-regulate. Hardcastle (2025) warns of an “atrophy of epistemic vigilance”: students risk losing the instinct to question sources and validate answers if they habitually outsource judgment. Early observations indicate that students who rely on AI often express lower confidence in their independent problem-solving (Anecdote: some studies of student attitudes).
  • Equity and Access: Generative AI may both democratize learning (by providing personalized tutoring) and widen gaps (favoring those with digital savvy). Some studies (not covered here) are exploring if AI can help low-performing students catch up; others note uneven adoption. This area needs more research (Evidence: Speculative).

5. Creativity, Writing, and Intellectual Formation

AI’s impact on creativity and writing is complex:

  • Creativity: Evidence suggests AI can amplify creative output, but typically only under certain conditions. Lu et al.’s field experiment found higher creativity ratings with ChatGPT but only when users engaged metacognitively. In contrast, uncontrolled AI use can homogenize work. Kosmyna et al. observed that essays written with AI were statistically more uniform and less original. Similarly, surveys indicate that habitual AI use tends to converge thinking around common templates (metacognitive trace: AI “amplifies the effect that social media did to groupthink” – expert comment). Thus, AI’s creative benefits may require active guidance. (Evidence: Moderate – one strong field study, others still emerging.)
  • Writing and Authorship: Students frequently use AI to draft, proofread, or even write entire assignments. Empirical findings raise concerns about “parasitic dependence”. For instance, after four months, Kosmyna’s LLM group “consistently underperformed at linguistic levels,” feeling less ownership and even failing to quote their own prior text. In contrast, participants who started brain-only then used AI showed improved prompts and retention. These patterns imply that AI can bypass the struggle that builds writing skill. However, when used as a collaborative scaffold (e.g. to generate ideas that a student then refines), AI can foster richer outlines and revisions. The distinction again hinges on student agency and feedback.
  • Intellectual Formation: Over time, a culture of passive AI use might erode “cognitive sovereignty”. Hardcastle (2025) argues that reliance on algorithmic knowledge sources shifts epistemology – students may stop being the arbiters of truth. Historical analogies (below) suggest each tech shift required educational adaptation: new curricula for calculators, etc. Without guidance, there is a risk that younger learners become “AI-displaced”, with gaps in critical reasoning (Hardcastle’s term). This is still prospective: evidence on long-term intellectual development is emerging (Evidence: Preliminary).

6. Augmentation vs. Replacement

The data highlight that context of use matters deeply. Across domains:

  • Structured Use (Augmentation): Studies find AI can augment cognition when users remain in control. For example, Kambanis (2026) describes a trial: when students first outlined their own answers before consulting an AI, they later demonstrated higher task engagement and reasoning, compared to students who used AI unstructured. The key was preserving human planning and critique. Similarly, Lu et al. (2025) emphasize that workers who actively guided AI (through reflection and iteration) reaped creative benefits. This aligns with cognitive science: offloading is powerful but only when aligned with one’s goals. Clark (2025) notes that ideally AI forms part of “well-regulated interaction” with our thinking.
  • Unguided Use (Displacement): By contrast, “free use” often means users slack off. In typical experiments, participants told to “use AI if you wish” tend to rely on it for core ideas and accept outputs uncritically. This leads to standard findings: offloading coupled with poor outcomes. For instance, in Kosmyna’s study, ChatGPT-led authors had uniformly weak performance; the group that switched off AI (LLM-to-Brain) still showed under-engagement. These patterns mirror past observations: anyone who “lets the machine do the heavy thinking” learns little.
  • Augmentative Scaffolds: Several researchers highlight design strategies to flip the balance. Recommendations include: requiring users to generate a plan or first draft themselves, limiting AI’s role (e.g. “edit only mode”), and training metacognitive skills. Early pilots of such curricula (e.g. prompts that stress validation of AI output) show promising effects on engagement (Education research in 2025–26, under review). These interventions are only beginning to be empirically tested (Evidence: Preliminary).

7. Social and Psychological Dimensions

  • Algorithmic Recommendation & Passivity: Beyond one-to-one interactions, the culture of AI mirrors that of social media. When content is algorithmically curated, users can develop “learned helplessness” – deferring decisions to opaque algorithms. Hardcastle notes that reliance on AI “quiets” students’ questioning habits. The broader “attention economy” compounds this: even modest AI assistance can incentivize acceptance of spoon-fed answers (no friction). The risk is a society less practiced at evaluating truth. While causal studies are scarce, analogous research on recommendation algorithms suggests that constant passive consumption correlates with lower reflective thinking.
  • Epistemic Dependence & Authority: “Algorithmic authority” means accepting AI-generated knowledge as if expert testimony. Students may become overconfident in AI outputs, unaware of biases in training data. This shift can subtly shape curricula and knowledge: as Hardcastle warns, “allow commercial training data… to shape what questions get asked and which methodologies appear valid.” In sum, social media analogues (filter bubbles, echo chambers) carry over: AI tools, if widely trusted, could entrench certain viewpoints under the guise of neutrality (Speculative but plausible).
  • Learned Helplessness & Cognitive Dependency: Psychologists caution that repeatedly deferring tasks to AI can train out skills. In a social-psych experiment, novices given hints on every step of a puzzle often never solved it alone later. Similarly, if students habitually ask “what to do” vs “how to think”, they may lose confidence. Rosenquist’s (1980) classic “Sisyphus effect” showed students given problem solutions performed worse on transfer tasks. Emerging surveys of heavy AI users report feelings of dependency and anxiety when AI is unavailable (some APA preprints). These are early findings, but suggest a risk of learned helplessness (Evidence: Preliminary).
  • Benefits to Motivation: On the positive side, some students find AI demystifies complex problems, encouraging experimentation (Anecdote: a pilot study on math homework). By reducing initial frustration, AI can sustain engagement for learners at risk of giving up. Again, this depends on pedagogy (Evidence: Speculative/Preliminary).

8. Historical Parallels

AI echoes past tech shifts. Socrates’ critique of writing has survived millennia as a cautionary tale about outsourcing memory. Later, fears that the Internet or calculators would “make us dumb” proved overblown: humans largely adapted. Some analogues:

  • Printing Press/Books: Once information became mass-produced, scholars worried students would not memorize facts. But literacy ultimately expanded knowledge (Scribner 1977).
  • Calculators (1980s): Educators feared mental arithmetic skills would vanish. In fact, curricula adjusted: rote calculation was deemphasized in favor of understanding concepts. Many argue overall math reasoning remained robust.
  • GPS (2000s): As noted above, studies (e.g. London taxi drivers vs bus drivers) show that navigating skills can atrophy if offloaded. However, in many societies people still navigate without tech. No collapse of spatial cognition has been documented, though there may be a cohort effect.
  • Internet Search (2010s): “Google Effect” studies found people remember the location of facts (i.e. the tool) rather than facts themselves. This reframed memory as knowing how to find information, a useful skill. Critical thinking tasks, however, still require discerning quality of sources – a skill that search engines did not directly teach.

These parallels suggest two lessons: 1) Adaptation is possible (new norms, curricula can evolve). 2) Tools often shift what skills matter. Just as calculators made memorizing multiplication tables less critical but raised emphasis on problem-solving, AI may similarly reshape educational priorities. Importantly, none of these shifts “destroyed cognition” altogether – they changed its distribution.

9. Counterarguments and Nuance (Against Pessimism)

While mounting evidence points to caution, there are strong counterpoints:

  • Evidence of Enhancement: Some rigorous studies show positive effects under controlled conditions. For instance, Iqbal et al. (2025) surveyed 465 student teachers and found that when AI use was mediated by good pedagogy (metacognitive strategies and collaborative reflection), cognitive offloading actually predicted higher academic achievement. In their SEM model, offloading and shared metacognition acted as mediators between AI usage and better learning outcomes. This suggests AI can free resources for deep learning, much like calculators free arithmetic effort for conceptual math.
  • Human-AI Collaboration: Experts like Clark (2025) argue against a simplistic doom view. He envisions AI as part of a new synergy: “all our stunning new resources should come together, creating a collaborative web that acts as a massive amplifier and transformer of creative human intelligence.” The analogy is that AI, like electricity, is a general-purpose amplifying force; it need not shrink minds. Indeed, as Clark notes, the brain evolved to exploit exactly this sort of external feedback loop. The key is fostering an “extended cognitive hygiene” – teaching people how to critically engage AI rather than blindly trust it.
  • Correlation vs. Causation: Many findings are correlational or short-term. E.g. Gerlich’s correlation between AI use and low thinking might reflect that less diligent students both rely on AI and score lower, rather than AI causing low skill. Similarly, the EEG studies measure only momentary engagement, not long-term ability. Critics argue (Kambanis 2026) that existing studies show immediate offloading during tasks but do not prove lasting cognitive decline. They caution that longitudinal studies on brain development are still lacking.
  • Counterexamples: Anecdotally, some educators report that properly guided AI use has improved student motivation and creativity, enabling students who were stuck to break through writer’s block or math hurdles. These stories are not yet in controlled trials, but they suggest the risk is not universal.

Synthesis: The weight of current data leans toward caution: many well-controlled experiments find immediate decreases in mental effort and recall. However, the evidence on long-term effects and conditional benefits is much weaker, as researchers acknowledge. The strongest positive evidence comes from carefully scaffolded use; without those structures, the likely outcome is partial cognitive atrophy, not wholesale stupidity. Thus, the debate is not settled: AI can both help and hinder thinking, depending on design and context.

10. Policy, Education, and Design Recommendations

Given the stakes, multiple strategies emerge from the literature:

  • Educational Strategies: Teach AI literacy, not just technical use. Emphasize critical engagement: For example, require students to generate initial responses before consulting AI, and to identify errors in AI output. Integrate exercises that build metacognition (self-reflection, fact-checking). Curricula should clarify the role of tools: calculators were once banned for tests until education adapted; similarly, open-book AI-inclusive exams might be needed. Encourage desirable difficulties – tasks that AI cannot fully handle without human insight.
  • Design Principles: Human-AI interfaces should promote transparency and accountability. For instance, AI writing assistants could highlight when content is AI-sourced (so users remain aware), or require human confirmatory steps. “Friction points” (e.g. quizzes about what the AI just gave you) could ensure engagement. Some suggest embedding explainable AI components that show reasoning steps. Interfaces that emulate Socratic questioning, rather than just supplying answers, might cultivate critical thinking. Researchers also call for AI to signal uncertainty and encourage users to verify sources.
  • Policy: At the institutional level, set guidelines for AI use in high-stakes learning (e.g. cap types of tasks allowed). Promote research into cognitive impact with funding. Consider regulations akin to mandating accuracy standards or registration for powerful models. More broadly, society might need safeguards against “algorithmic authority”: encourage media literacy to question AI content, and maintain diverse knowledge sources (libraries, experts). Academic and workplace assessments may need redesign: focus on tasks AI cannot easily substitute (high-level synthesis, originality).
  • Continued Research: Support longitudinal studies on AI’s cognitive effects (e.g. multi-year tracking of student cohorts). Interdisciplinary work (neuroscience, ed psych, HCI) should probe questions like: Does early AI exposure shift developmental trajectories? Are there “critical periods” of cognitive imprinting? Likewise, study which pedagogical interventions best leverage AI without cost. Several authors stress the importance of “outside-view” analyses – essentially, using broad base rates from past tech shifts to contextualize hypotheses about AI’s impact.

11. Key Findings

  • Cognitive Offloading: Delegating thought tasks to AI is a modern extension of an ancient concept. It frees mental resources but risks shallow processing if overused【52†L384-L388】.
  • Distinctive Features of AI: LLMs generate information, not just retrieve it. They can hallucinate (unlike calculators) and have no fixed domain, which blurs responsibility for thinking.
  • Memory & Effort: Empirical studies (EEG, writing tasks) consistently show AI use reduces cognitive load and recall: e.g. ChatGPT-assisted writing yielded 55% weaker neural connectivity and 32% lower cognitive load than solo writing. Memory recall suffered accordingly (83% of AI users couldn’t recall their own text). (Strong Evidence)
  • Critical Thinking: Survey data link higher AI use with lower critical thinking scores (Gerlich 2025: r≈–0.5). Lab studies find AI users tend to accept answers without scrutiny. This suggests a cognitive cost to offloading thinking. (Moderate Evidence)
  • Creativity & Writing: AI can speed writing (30% faster in one study) and help generate ideas, but also cause “cognitive fixation” and diminished ownership. Field research found ChatGPT increased creativity scores only for those with strong self-regulation (metacognition). (Moderate Evidence)
  • Educational Outcomes: AI aids (e.g. tutors, copiers) have mixed effects. Large studies of students show efficiency gains but no quality gains: faster task completion without better scores. Preliminary evidence suggests adaptive use of AI (paired with pedagogy) can enhance learning, whereas misuse undermines it.
  • Neuroscience: Early EEG/fMRI work (Kosmyna et al. 2025) reveals that AI-assisted task performance involves weaker neural engagement across language and associative networks. The “brain-only” condition consistently showed richer connectivity. (Strong Evidence for short-term tasks)
  • Social Dependency: Experts warn of a shift to “algorithmic authority” where users defer too much to AI. Observational data (Anthropic, EdTech) already show students frequently using AI to solve homework, raising concerns about losing skills to verify knowledge independently.
  • Counterarguments: AI is not inherently “mind-numbing” if used well. It can be a cognitive amplifier. Structured, critical use of AI yields better outcomes. Past fears (e.g. about calculators) were partly allayed by educational adaptation; similar adaptation may be key here. Importantly, causality is not fully established – much of the data show correlation or short-term effects. Long-term cognitive trajectory studies are not yet available.

12. Open Questions and Future Research

  • Longitudinal Impact: Do children raised with AI as an “intellectual prosthetic” develop different cognitive patterns by adulthood? Are there critical windows (e.g. early literacy vs later)?
  • Societal Effects: How will reliance on AI for news, analysis, and conversation affect collective reasoning and democracy? Does algorithmic authority concentrate knowledge in the hands of a few (Big Tech), potentially biasing epistemic norms?
  • Individual Differences: Who is most at risk (or benefit)? Do high-IQ or highly motivated individuals adapt by using AI as a tool, while others become passive? Early data suggest education level moderates outcomes.
  • Skill Trade-offs: Which cognitive skills decline vs. improve with AI? For example, is rote memory decaying while pattern synthesis or interdisciplinary thinking improve?
  • Interventions Efficacy: Which pedagogical or design interventions most effectively maintain engagement? Field trials are needed (e.g. compare class sections with different AI scaffolds).
  • AI Literacy and Metacognition: How best to teach “AI literacy”? How to scale metacognitive training (so users know how to question and use AI)?
  • Metrics and Measurement: Developing reliable measures of “offloading” in naturalistic settings is challenging. Can we track neural or behavioral signs of dependency in everyday life (beyond lab essays)?

13. Implications and Recommendations

  • For Education: Curricula must evolve. Integrate AI as a standard tool (like calculators): teach when and how to use it, but also ensure core skills are practiced. Use AI bans judiciously (e.g. ban it only until concept is learned). Emphasize AI critique skills in lessons.
  • For Journalism/Art: Creativity and reporting that require human insight (e.g. empathy, context awareness) should be highlighted. Fact-checking protocols become more vital. Artists/journalists may emphasize authenticity and human voice as counters.
  • For Democracy: Civic education should include understanding AI’s role in information dissemination. Platforms might label AI-generated content. Civic tools could focus on crowd-sourced verification to counter algorithmic filtering.
  • Practical Tips for Users: Individuals should adopt “think-first” habits: e.g. attempt problems in private before seeking AI help; question AI answers; use AI for idea-generation, not final answers; maintain diversified offline habits (memory exercises, brainstorming on paper).
  • Designers & Policymakers: Encourage AI designs that nudge user engagement (e.g. periodic quizzes, mandatory source citations). Consider rights-based approaches: perhaps the idea of “cognitive sovereignty” – the right to reliable knowledge. Governments might fund public AI literacy campaigns, and research on cognitive effects.

References

Explore more on these topics:

Mindset 💡Your art career starts in your head

Interview 🎤 Behind the scene

Art Grants ★ What gets you funded

Comments

0
Poem Johnson Avatar

Pro tips: Ask questions, Request feedback, share works-in-progress

Leave a Reply

Your email address will not be published. Required fields are marked *