The fixer’s dilemma: Chris Lehane and OpenAI’s impossible mission

Bad news in Chris Leh is one of the best businesses in Making. In the years of Clinton, Al Gore’s Press Secretary, AirBNB’s Chief Crisis Manager, knows how to spin in Lehan – every regulatory nightmare from the Brussels to Brussels. Now he is in two years that he could be the most impossible gig: Opena’s global policy VP, his work is to explain to the world that Opina curses a democraticization of artificial intelligence when the company behaves increasingly, well, every technology that is claimed to be different.

I had 20 minutes on stage with him Developed Conference in Toronto early this week – 20 minutes to cross the talking points and the actual conflicts eaten in the image created with the opening caution. It was not easy or completely successful. Lehan is really good at his job. He is desirable. He listens reasonable. He acknowledges uncertainty. He even talked about waking up in the morning at three o’clock in the morning that none of these would actually benefit humanity.

When your company is sub -poening of critics, the economically frustrated cities of water and electricity are drying up and making the dead celebrities lived to emphasize your market domination is not too good.

The company’s Sora problem is really at the root of everything else. Video generation equipment was launched last week with a seemingly baked with copyrighted components. It was already a bold step for a company to sue the New York Times, Toronto Star and half the publishing industry. It was also brilliant from a business and marketing point of view. Invite-Cable App As the app is grown At the top of the App Store As people have created their own digital version, OpenAI CEO Sam Altman; Characters like Pikachu, Mario and Cartman of “South Park”; And dead celebrities like Tupak Shakur.

With these characters, the decision to launch this new version of Sore has been asked about what the decision to launch is, Lehan gave me the standard pitch: Sora is a “general purpose technology” like the power or printing press, the democraticization of creativity for the people without talent. He even can make a video of a self-covered creative zero-he said on stage.

The one he dancing around is the openly “let” rights holders get their jobs from using Sora to train, which usually does not work using copyright. Then, people who really liked using copyrighted images after the opeina noticed, “developed” on one side Opt-in modelThe This is not actually a repeat. Checking how far you can go. (And anyway, though the motion Picture Association Did some noise Last week about legal threats, Opina seems to have moved away with a lot))

Naturally, the situation remembers the growing publishers who accuse the opening of the openings for training on their jobs without sharing financial plunders. When I pressed on the lean of the publishers to get separated from the economy, he called for a fair use, which is American legal doctrine, which is considered to be the rights of Creator against the public access to knowledge. He called it a secret weapon of US tech domination.

May be. However I used to do recently Al -Gore’s interview – Lehan’s Old Boss – And understand that anyone can just ask Chatzipt about it instead of reading my piece on TechCrunch. “It’s ‘Recruit’,” I said, “but it is also a replacement” “

For the first time, Lehan dropped his spill. “We all need to get it out,” he said. “Sitting on the stage here and our new economic earning models need to be found and it’s really easy. But I think we will.” (In short we are making it as soon as we make it))

Then there are infrastructure questions that no one wants to answer honestly. OpenAI is already operating a data center campus in Abelin, Texas, and recently grounded with Oracle and Softbank to a huge data center at Ohio Lordstown. Lehan has compared to the advent of electricity with AI’s accessibility-they say they are still playing catch-ups-but the Stargate project of Openai is probably aiming for the same economically challenged places to set up their huge hunger for water and electricity.

We were asked while sitting on the bill that these communities would only benefit from the bill or just to get this bill, lehan went into gigawats and geopolitics. Opeena needed about one gigawatt strength every week, he mentioned. China came up with 450 gigwatts and 33 nuclear benefits last year. If democracy wants a democratic AI, they have to compete. “I am hopeful that it will modernize our energy systems,” he said, drawing a picture of a re -artistic American with the converted power grids.

It was inspirational. But Lordstown and Abelin’s people were not answering whether their utility bills were going to see spikes when Openai John F. Kennedy and the infamous big video made (video generation. Maximum energy-intensive AI Outside.)

Which brought me to my most uncomfortable example. The Jewda Williams spent the day before we begged our interview to stop sending strangers to strangers in our interview. “You’re not creating art“He wrote.” You are creating hotdogs, excessively processed hotdogs from people’s lives. “

When I asked how the company met with its mission with such intimate damage, Lehan’s responsible design, test framework and processes, including government partnerships, answered. “There is no playbook for this stuff, isn’t it?”

Lehan showed weakness in some moments, he said that he woke up at 3am every night and was concerned about democracy, geo -politics and infrastructure. “There is a lot of responsibility with it.”

If these moments were designed for the audience, I believe him. In fact, I left Toronto, thinking I saw a master class in political messages – while throwing questions about the company’s decisions while making an impossible needle threading, I know that he even did not agree. Then it happened on Friday.

Nathan Calvin, a lawyer working in AI Policy in the Encode AI, revealed that I was talking to Lehan in Toronto as well as OpenAA. Sheriff’s deputy in his house In Washington, DC, to serve him a subpoena during dinner. They sought his personal message with California MLA, college students and former OpenAI staff.

Calvin is accusing the Opeena of intimidating the new part of the SB3 of California. He says that the company had armed with his legal war with Elon Kasturi as an excuse to target critics, which was secretly financed by musk. In fact, Calvin says that he fought opposition to the Opena against an AI protection bill in California, and when he found that the company claimed “it worked to improve”, he “literally laughed loudly.” In a social media, he called Lehan, especially “the master of political dark art” in Lehan.

In Washington, it can be a praise. In an organization like Openai, the mission whose mission is “creating AI for the benefit of all humanity”, it sounds like a complaint.

The more important thing is that Openai’s own people are opposed to what they are becoming.

The highest as my peer Been reported last weekAfter the release of Sora 2, several current and former employees went to social media, expressed their confusion, with Boz Barack, with an OpenAI researcher and professor of Harvard who, who Wrote about Sora 2 This is “technically amazing but prematurely congratulate yourself to avoid other social media apps and dipfek problems.”

On Friday, Josh Achium – Opehena’s Mission Migration Head – Calvin’s allegations tweeted some more significant. They prefacking his comments that “probably my whole career is at risk” wrote: “We cannot do things that turn us into a frightening force instead of a virtue. We have a duty for all humanity and a mission is significantly high.

It The .Soming. An OpenAI executive is publicly asking whether his company is becoming “a scary force instead of virtue,” is not equal to the competitor’s shot or asking questions to a reporter. This is the one who chosen to work in Open, who is a believer in his mission and who is now acknowledging the crisis of conscience despite the professional risk.

This is a crystal moment. You can be the best political operating of the tech, a master in navigating the impossible circumstances and still finished work for an organization whose activities are increasingly opposed to its described values ​​- conflicts that can only be intense as an open race for artificial general intelligence.

This is my wondering if the real question is not whether Chris Lehan can sell the Openai mission. It is others – critically, the other people working there – whether it still believes it is.