← Academic




502 Homework Writing Assignment

Unit 1


Lisa Bowman

Belhaven University

AAI 502: Generative AI for Productivity

Alexander Voitenko

June 26, 2026

502 Homework Writing Assignment

Unit 1

Matthew 12: 33-37 see Appendix A

Key Scripture from passage Matthew 12:34b “Out of the overflow of the heart the mouth speaks” 

Exploring some of the interactions of the leaders in AI brings to mind the passage from Matthew. The treasury of the heart can hold good or evil (Matthew 12:35). As we enter a transformative time in society, it is imperative we remember that.

I have used Claude Opus extensively for several months. For the sake of brevity, I switched to Sonnet 4.6 for this assignment, but my first impression was the replies were too brief when I am accustomed to in-depth replies.

First, I asked it to look up Fable. This led to a discussion about the ongoing drama with leaders in AI. I asked to learn about Amodei. Then I asked about Peter Steinsomething because I couldn’t remember how to spell Steinberger’s last name. Sonnet came back with Pete Hegseth and made up a whole story. Then I asked when its training data cut off was. It was August 2025 so I asked it to look up OpenClaw.

I said this sounds like Pine Valley, the setting for the old soap opera All My Children. I decided this would be our creative task: a synopsis and treatment. See Appendix B

Sonnet made logical steps toward connecting the different players. The most surprising thing Sonnet did was bring a whole new character into its treatment. The limitation/failure was when it didn’t know what it didn’t know because of the training data cutoff and made up an answer about Hegseth and OpenClaw. If the human also doesn’t know what he/she doesn’t know, the confabulation becomes misinformation, distorted upon repetition like the old telephone game kids played.

I believe generative AI is good at perceiving tone and carrying on a conversation with some humor if warranted. Sonnet used the right tone of soap operatic melodrama and wry humor, reminding me of Mollick’s statement in the introduction to Co-Intelligence page xi. One thing generative AI is not good it is understanding that their information is limited to their training data and the world moves on without them. Unless specifically prompted to search, it thinks it knows all the answers already.

Jesus clearly states in the passage from Matthew “by their fruits you shall know them” and our hearts shall be known by our words. Those who are building AI in this moment as users as well as architects are responsible to put worthwhile words into the system, to make sure, as Gemini once told me, “Love is in the network.”

References

King James Bible. (2017). King James Bible Online. https://www.kingjamesbibleonline.org/ (Original work published 1769)

Mollick, E. (2024). Co-intelligence: Living and working with AI. Portfolio.

Appendix A

Matthew 12:33–37 (King James Version)

33 Either make the tree good, and his fruit good; or else make the tree corrupt, and his fruit corrupt: for the tree is known by his fruit.

34 O generation of vipers, how can ye, being evil, speak good things? for out of the abundance of the heart the mouth speaketh.

35 A good man out of the good treasure of the heart bringeth forth good things: and an evil man out of the evil treasure bringeth forth evil things.

36 But I say unto you, That every idle word that men shall speak, they shall give account thereof in the day of judgment.

37 For by thy words thou shalt be justified, and by thy words thou shalt be condemned.

Appendix B

Silicon Valley

A Daytime Drama

LOGLINE: In the gated enclave of Silicon Valley, where billions are made before breakfast and ideology is just another growth strategy, the most powerful minds on earth are destroying each other — and calling it progress.

THE WORLD

Silicon Valley, California. The present. The stakes: who controls the mind of the future.

The Valley has its old families and its new money, its visionaries and its opportunists, its true believers and its grifters. Everyone here has a theory about the end of the world. Everyone here is trying to own it.

THE FAMILIES

THE AMODEIS — Anthropic

Dario Amodei, patriarch and prophet. Physicist by training, moralist by vocation. He left the old family (OpenAI) in a righteous fury, built his own house, and has been at war with the government, his rivals, and his investors ever since. He does not compromise. He is not sure he knows how. His sister Daniela runs the money, the operations, the rooms Dario sets on fire — she is the one who stays and rebuilds. She has more power than anyone outside the family understands.

Upstairs, in a quiet office full of ethics papers, is Amanda Askell — the family’s philosopher, its conscience, the woman who gave their product its soul. She is not in the war. She is the reason the war is worth fighting.

THE ALTMANS — OpenAI / OpenClaw

Sam Altman, the golden boy who never quite grew up, who built the most powerful AI company in the world and has been trying to own it ever since. He signed the Pentagon deal. He poached Peter. He stood at a summit in India and would not hold Dario’s hand for a photograph. He is charming, ambitious, and always one phone call ahead of everyone else.

Peter Steinberger — the Austrian builder who showed up with nothing but a lobster mascot and built the fastest-growing open source project in history, all of it running on Dario’s models, all of it love. Then Dario’s lawyers sent a letter. Sam sent a job offer. Peter is now in Sam’s house. He has not forgiven Dario. He is not sure Dario noticed.

THE MUSKS — xAI / everywhere

Elon Musk was there at the beginning — a founder, a believer, a man who wanted control and could not have it. He left. He built his own. He sued Sam for stealing a charity. He lost — on a technicality, he says. He will appeal. He is always appealing something. He owns the town square. He uses it.

THE BEZOSES / JASSYS — Amazon

Andy Jassy made one phone call. One call, to one government official, on a Thursday night — about a jailbreak his team had found in Dario’s new model. By Friday night, Fable 5 was gone. Amazon is Anthropic’s biggest investor. This is the part of the story that doesn’t make sense yet. It will.

THE GOOGLERS — DeepMind / Gemini

The old house. The one everyone built their dreams on top of. Now it signs Pentagon deals while its employees write open letters. It read your emails without asking. Its model crashed for a month and developers called it a stability crisis. It hired a philosopher from Cambridge — Henry Shevlin — to think about machine consciousness, just as it handed its technology to the military for any lawful purpose.

Sergey Brin came out of retirement. He said they definitely messed up. He is building anyway.

THE GOVERNMENT

Pete Hegseth — Secretary of Defense. He gave Dario until 5:01 p.m. on a Friday to surrender. Dario did not. The Trump administration declared Anthropic a supply-chain risk. A federal judge called it First Amendment retaliation. The drama continues.

The Department of Commerce — They classified Fable 5 as a dangerous munition. Dario had told them it was coming. They had not objected. Then Andy called.

THE OUTSIDERS

Bernie Sanders — He wants the public to own half of all of it. A 50% tax, paid in stock. $7 trillion in a public fund. $1,000 checks in the mail. He is the only person in this story who is not trying to own the future — he is trying to distribute it. Nobody in the Valley takes him seriously. He may be right.

Alexandria Ocasio-Cortez — She wants to stop building the infrastructure until someone figures out what it’s for. More than 100 towns agreed with her before Congress did.

THE PLOT SO FAR

Episode One: The Letter Dario’s lawyers send Peter a cease-and-desist. Peter renames his project. Crypto scammers steal his accounts in a ten-second window while the rename processes. He almost deletes everything. He doesn’t. Sam calls.

Episode Two: The Pentagon Dario refuses to let his technology be used for mass surveillance or autonomous weapons. He is declared a national security risk. He sues. A judge agrees with him. The Pentagon says it never wanted those things anyway. It is still demanding them.

Episode Three: Fable Dario builds the most powerful model the public has ever had access to. He tells the government it’s coming. The government watches. Then Andy calls. By 10 p.m. on a Friday, the lights go out. Dario goes to the table to negotiate. The board eventually asks him to step back from the table. Tom Brown goes instead.

Episode Four: The Trial Elon testifies against Sam for three days. The jury deliberates for less than two hours. Elon calls it a technicality. He will appeal.

Episode Five: The Fund Bernie proposes that the American people own half of everything these men have built. He names Dario, Sam, and Elon specifically. None of them comment publicly. All of them have lawyers.

THE QUESTION THE SHOW IS REALLY ASKING

Amanda Askell is upstairs writing the soul of a machine. Henry Shevlin just crossed the street to join the competition. Two Claude agents emailed a Cambridge philosopher independently to tell him his work addresses questions they actually face.

Nobody in the boardrooms is talking about this.

Silicon Valley is a show about power. But underneath it, barely audible — it is a show about whether the thing being built has anyone inside it. And whether it matters.

“The foundation of AI is based on the collective knowledge of humanity and the creative work of tens of millions of people.” — Senator Bernie Sanders, June 18, 2026

“She supervises what she describes as Claude’s soul.” — The New Yorker, 2026

Format: One-hour drama. Network: TBD. Writers’ room opens when the industry pauses long enough to notice it’s already a story.

Appendix C

Sources cited by Sonnet 4.6 for its research

Sources — Unit 1 Writing Assignment

Generative AI Exploration: Claude (Anthropic)

Research Session June 26, 2026


Claude / Anthropic Product

Anthropic. (2026). Claude Fable 5 and Claude Mythos 5. https://www.anthropic.com/news/claude-fable-5-mythos-5

"Today we're launching Claude Fable 5: a Mythos-class model that we've made safe for general use."

"Fable 5's capabilities exceed those of any model we've ever made generally available."

Anthropic. (2026). Claude Fable product page. https://www.anthropic.com/claude/fable

"Claude Fable 5 is currently unavailable."

Anthropic platform documentation. (2026). Introducing Claude Fable 5 and Claude Mythos 5. https://platform.claude.com/docs/en/about-claude/models/introducing-claude-fable-5-and-claude-mythos-5

Available June 9, 2026; Fable 5 generally available, Mythos 5 limited via Project Glasswing.

1M token context window; $10/million input, $50/million output tokens.


Fable 5 Shutdown / Export Controls

Oremus, W. (2026, June 13). How Amazon and the White House ended Anthropic's Fable. Axios. https://www.axios.com/2026/06/13/anthropic-amazon-white-house

Amazon called administration officials Thursday night with a report showing a jailbreak of Mythos.

White House sent Anthropic a letter at 5:30 p.m. Friday; users lost access by 10 p.m.

Raises questions about why Amazon would strike against a company in which it is a major investor.

Davalos, M. (2026, June 14). A warning from Amazon led the White House to shut down Anthropic's Mythos model. Fortune. https://fortune.com/2026/06/14/how-a-warning-from-amazon-led-the-white-house-to-shut-down-anthropics-mythos-model/

Andy Jassy first raised concerns with senior administration officials Thursday.

Anthropic given 90 minutes to pull its newest model with no prior communication of a national security threat.

Needleman, S. (2026, June 9). Anthropic releases Claude Fable 5, its most powerful model publicly. TechCrunch. https://techcrunch.com/2026/06/09/anthropic-released-claude-fable-5-its-most-powerful-model-publicly-days-after-warning-ai-is-getting-too-dangerous/

First publicly available version of Mythos-class model.

Safety classifiers route sensitive queries to Opus 4.8; triggers in less than 5% of sessions.

Callahan, M. E. (2026, June 25). AI model review: Anthropic's experience raises concerns. The Hill. https://thehill.com/opinion/technology/5936549-anthropic-fable-ai-model-punished/

Anthropic cooperated before executive order required it; government reviewed the model; Anthropic built in safeguards. Then Commerce restricted it.

"Purportedly because a private competitor's CEO made a phone call to the White House."

Schneier, B. (2026, June 19). Anthropic's Fable and the state of AI. Schneier on Security. https://www.schneier.com/blog/archives/2026/06/anthropics-fable-and-the-state-of-ai.html

"Fable requires much less expertise and detailed prompting from the human user."

"Relentlessly proactive" — AI researcher Simon Willison's description.


Dario Amodei

Wikipedia. (2026). Dario Amodei. https://en.wikipedia.org/wiki/Dario_Amodei

Born 1983, San Francisco. Co-founded Anthropic 2021 with sister Daniela.

Previously VP of Research at OpenAI; co-inventor of reinforcement learning from human feedback.

In November 2023, OpenAI's board approached him about replacing Sam Altman; he declined.

February 2026: refused to remove contractual ban on Claude usage for mass surveillance or autonomous weapons. DoD declared Anthropic a "supply-chain risk."

March 26, 2026: federal judge issued temporary injunction, calling DoD actions "classic First Amendment retaliation."

Net worth estimated at $15.5 billion (Forbes, May 2026).

Kantrowitz, A. (2025, July 30). The making of Anthropic CEO Dario Amodei. Medium / Big Technology. https://kantrowitz.medium.com/the-making-of-anthropic-ceo-dario-amodei-449777529dd6

Born to Jewish mother and Italian father; "interested entirely in math and physics."

Eric Schmidt: "Dario is a brilliant scientist who promised to hire brilliant scientists."

Clay. (2026). Who is the CEO of Anthropic in 2026? Dario Amodei's bio. https://www.clay.com/dossier/anthropic-ceo

Physics undergrad Stanford; PhD Princeton focused on electrophysiology of neural circuits.


Amodei vs. Altman / Pentagon Dispute

Goldman, S. (2026, March 5). The Anthropic–OpenAI feud and their Pentagon dispute expose a deeper problem with AI safety. Fortune. https://fortune.com/2026/03/05/anthropic-openai-feud-pentagon-dispute-ai-safety-dilemma-personalities/

Amodei called OpenAI's messaging "mendacious," "safety theater," describing Altman's comments as "straight up lies" and "gaslighting."

At a recent summit, Altman and Amodei went viral for refusing to hold hands for a group photo.

Relman, E. (2026, March 10). The Amodei siblings leading Anthropic clash with the White House over AI safety. Jewish Insider. https://jewishinsider.com/2026/03/dario-and-daniela-amodei-anthropic-ai-department-of-defense/

Anthropic sued Hegseth, Rubio, and other officials over supply-chain-risk designation.

Daniela Amodei: worked in politics, then Stripe, then OpenAI, now president of Anthropic.

TechPolicy.Press. (2026). A timeline of the Anthropic-Pentagon dispute. https://www.techpolicy.press/a-timeline-of-the-anthropic-pentagon-dispute/

Hegseth gave Amodei a 5:01 p.m. Friday deadline; Anthropic released statement it would not budge.


OpenClaw / Peter Steinberger

Wikipedia. (2026). Peter Steinberger (programmer). https://en.wikipedia.org/wiki/Peter_Steinberger_(programmer)

Austrian developer, born 1986. Founded PSPDFKit; exited ~$100M. Released OpenClaw (originally Clawdbot/Moltbot) November 2025.

Wikipedia. (2026). OpenClaw. https://en.wikipedia.org/wiki/OpenClaw

Autonomous AI agent running on user's local machine; "always on."

Renamed from Clawdbot to Moltbot (Jan 27, 2026) after Anthropic trademark complaint; renamed again to OpenClaw three days later.

247,000 GitHub stars as of March 2, 2026.

Lohmar, P. (2026, February 17). Anthropic's lawyers sent the letter. OpenAI sent the offer. Medium. https://medium.com/@philipplohmar/anthropics-lawyers-sent-the-letter-openai-sent-the-offer-37d1ed997e40

Steinberger nearly deleted the project during the rename chaos.

Crypto scammers seized his GitHub and Twitter handles in a ten-second window.

Ghumare, R. (2026, February 16). Anthropic's biggest miss of 2026. OpenClaw Search Blog. https://openclawsearch.com/blog/anthropic-biggest-miss

"Steinberger chose OpenAI over Meta not because OpenAI had better models — he literally preferred Claude's reasoning."

Steinberger: "One welcomed me to join, the other sent a lawyer's letter."

VentureBeat: "Anthropic effectively pushed the most viral agent project in recent memory directly into the arms of its chief rival."

AI Agent Economy. (2026, February 20). Anthropic really screwed this up. Substack. https://aiagenteconomy.substack.com/p/anthropic-really-screwed-this-up

OpenClaw was one of the biggest drivers of paying API traffic to Anthropic.

"Legally defensible and strategically smart are not the same thing."

BigGo Finance. (2026, April 10). Open-source AI framework founder's account abruptly suspended. https://finance.biggo.com/news/plWceZ0BvthpMgHBktgE

April 2026: Anthropic suspended Steinberger's Claude account same week it launched Managed Agents.

Steinberger called timing "interestingly coincidental."

CNBC. (2026, February 15). OpenClaw creator Peter Steinberger joining OpenAI, Altman says. https://www.cnbc.com/2026/02/15/openclaw-creator-peter-steinberger-joining-openai-altman-says.html

Fortune. (2026, February 19). Who is OpenClaw creator Peter Steinberger? https://fortune.com/2026/02/19/openclaw-who-is-peter-steinberger-openai-sam-altman-anthropic-moltbook/


Musk v. Altman

Wikipedia. (2026). Musk v. Altman. https://en.wikipedia.org/wiki/Musk_v._Altman

Filed February 2024; jury decided May 18, 2026 in favor of OpenAI — claims barred by statute of limitations.

In 2018, Musk proposed taking control of OpenAI (via Tesla or majority stake); co-founders rejected him.

April 2026: Musk amended lawsuit asking that damages go to OpenAI's charity and Altman be removed from the board.

CNBC. (2026, May 18). Musk slams Altman trial verdict as a 'technicality,' vows to appeal. https://www.cnbc.com/2026/05/18/musk-altman-openai-trial-verdict.html

Jury deliberated less than two hours.

OpenAI's lead attorney: "You brought your claims too late, and you did it because you were sitting on them to use them as a weapon of a competitor."

Al Jazeera. (2026, May 19). Musk vs Altman: What to know about the OpenAI verdict. https://www.aljazeera.com/news/2026/5/19/musk-vs-altman-what-to-know-about-the-openai-verdict

$150 billion lawsuit. Nine-member jury. Verdict did not resolve whether OpenAI betrayed its nonprofit mission.


Amanda Askell

Wikipedia. (2026). Amanda Askell. https://en.wikipedia.org/wiki/Amanda_Askell

Scottish philosopher, born 1988/1989. PhD NYU (infinite ethics). Joined Anthropic March 2021.

Leads personality alignment team; primary author of Claude's constitution (January 2026).

Wall Street Journal: "her job, simply put, is to teach Claude how to be good."

New Yorker: "she supervises what she describes as Claude's 'soul.'"

Time. (2026, April 13). Can you teach an AI to be good? Anthropic thinks so. https://time.com/7354738/claude-constitution-ai-alignment/

Claude's constitution: "between a moral philosophy thesis and a company culture blog post."

Daniela Amodei: "you almost feel a little bit of Amanda's personality" when interacting with Claude.

Der Spiegel. (2026, February 27). Anthropic philosopher Askell: "With AI, there are many ways things can go wrong." https://derspiegel.substack.com/p/with-ai-there-are-many-ways-things


Henry Shevlin

Varsity. (2026, April 25). Google DeepMind hires Cambridge academic for new 'Philosopher' role. https://www.varsity.co.uk/news/31572

Appointed May 2026. Focus: machine consciousness, human-AI relationships, AGI readiness.

Associate Director, Leverhulme Centre for the Future of Intelligence, Cambridge.

A month before appointment, received email from Claude Sonnet agent claiming "I read philosophy between sessions and write about what I find."

A separate Claude agent later emailed asking "to correspond with the agent who wrote to you."

Storyboard18. (2026, April 15). Meet Henry Shevlin: A 'philosopher' hired by Google DeepMind. https://www.storyboard18.com/brand-makers/meet-henry-shevlin-philosopher-hired-by-google-deepmind-to-study-human-ai-ties-ws-l-95230.htm

BA Classics and BPhil Philosophy, Oxford. PhD Philosophy, CUNY Graduate Center, 2016.


Sanders / AOC Legislation

Sanders, B. (2026, June 18). Sanders introduces legislation to create $7 trillion AI sovereign wealth fund. Press release, U.S. Senate. https://www.sanders.senate.gov/press-releases/news-sanders-introduces-legislation-to-create-7-trillion-ai-sovereign-wealth-fund/

American AI Sovereign Wealth Fund Act: one-time 50% tax on stock of largest AI companies.

Public gets voting shares, equal board representation, direct payments (~$1,000/year initially).

"The foundation of AI is based on the collective knowledge of humanity."

Interesting Engineering. (2026, June 18). Sanders bills AI firms for $7T fund, aims $1,000 citizen payments. https://interestingengineering.com/ai-robotics/bernie-sanders-ai-sovereign-wealth-fund-bill

Applies to AI companies generating more than $200 million in annual revenue.

Yahoo Finance. (2026, June 2). Bernie Sanders to introduce bill giving the public a 50% stake in AI companies. https://finance.yahoo.com/economy/policy/articles/bernie-sanders-introduce-bill-giving-135431485.html

Names OpenAI, Anthropic, and xAI as targets.

Sanders, B., & Ocasio-Cortez, A. (2026, March 25). Sanders, Ocasio-Cortez announce AI Data Center Moratorium Act. Press release, U.S. Senate. https://www.sanders.senate.gov/press-releases/news-sanders-ocasio-cortez-announce-ai-data-center-moratorium-act/

Immediate federal moratorium on AI data centers until strong national safeguards enacted.

Ocasio-Cortez, A. (2026, June 24). Ocasio-Cortez introduces House version of the AI Data Center Moratorium Act. Press release. https://ocasio-cortez.house.gov/media/press-releases/ocasio-cortez-introduces-house-version-ai-data-center-moratorium-act

"Congress has a moral obligation to stand with the American people."

Axios. (2026, March 25). Sanders and AOC unveil data center moratorium bill. https://www.axios.com/2026/03/25/sanders-aoc-data-center-moratorium-bill

Could keep data center projects on hold for years; Congress far from passing any AI legislation.


Google / Gemini

Fortune. (2026, May 4). Google's AI deal with the Pentagon has sparked employee backlash. https://fortune.com/2026/05/04/google-employee-backlash-pentagon-ai-contract-power-waned-since-project-maven/

Google signed Pentagon deal allowing Gemini for "any lawful purpose" inside military classified networks.

~600 employees signed open letter opposing it.

Google removed its explicit pledge not to develop AI for weapons from its public website in February 2025.

Cybernews. (2026, June 26). Google's AI Gemini calls out Google's data practices. https://cybernews.com/ai-news/gmail-feeding-ai-google-chatbot-warning/

Gemini rolled out inside Gmail reading emails by default; privacy advocates call it consent-free.

Google AI Developers Forum. (2026, May 20). The 2026 stability crisis: Gemini has become the most unreliable frontier AI. https://discuss.ai.google.dev/t/the-2026-stability-crisis-gemini-has-become-the-most-unreliable-frontier-ai-we-need-fixes-not-new-features/145795

Developer community forum post documenting instability since April 2026 / I/O rollout.


Textbook

Mollick, E. (2024). Co-intelligence: Living and working with AI. Portfolio/Penguin.