Three big ChatGPT developments that undermined trust in OpenAI

I thought the OpenAI GPT-4o demo was great. The new model Chat GPT A significant upgrade over competitors. Google demoed a similar multimodal assistant feature with Gemini, but that demo wasn’t live and the feature won’t be available until later this year. The GPT-4o upgrade was rolled out shortly after a short press event.

Apple OpenAI I was working on ChatGPT to be integrated into upcoming iOS 18 update for iPhone It meant a lot.

You’d think OpenAI wouldn’t do anything to ruin the demo, but since then we’ve seen three troubling developments that call into question our faith in the company: OpenAI isn’t as old as the big tech companies that many hate, but it’s not a startup either, and by all appearances, it’s already starting to act like a big tech company.

Ilya Sutskever’s departure

Chronologically, the first thing that happened after the GPT-4o update was Ilya Sutskever leaves OpenAIIf you follow the inner workings of technology companies like OpenAI, it won’t necessarily come as a surprise to hear that the talented people behind the ChatGPT technology are leaving the company.

Sutskever may have played a major role in Sam Altman’s brief resignation last fall, and also supported Altman’s return as CEO. What actually happened there remains unclear, but Sutskever has essentially disappeared from public view. After a tumultuous board meeting in November, his departure seemed imminent, but it’s also worrying.

Many saw Sutskever as the main check on OpenAI, ensuring that it developed safe AGI – artificial general intelligence that was in the interest of humans and wouldn’t necessarily lead to the downfall of humanity.

OpenAI’s superalignment team collapses

After Sutskever left the company, the situation became even worse. Jan Reicke, co-leader of OpenAI’s superalignment team, said: left In a series of tweets, Reike said the team was struggling to secure enough resources to ensure the secure development of ChatGPT.

He said he had reached a “breaking point” after disagreements with OpenAI executives “about the company’s core priorities.”

“I believe we need to devote more resources to preparing for the next generation model, including security, surveillance, preparedness, safety, adversarial resistance, (super)alignment, confidentiality, societal impact, and related topics,” Reike said. “These problems are very hard to get right, and I’m concerned that we’re not on track to get there.”

Rijke is the latest player to leave the team in recent months.

In the two reports, Vox detail In addition to the departure of OpenAI’s SuperAlignment team, OpenAI employees Speaking to the press After leaving the company.

Does this mean OpenAI is about to unleash AGI on the world without the proper guardrails? Probably not, but it does mean the company is potentially moving much faster and more dangerously than Sutskever and Leike would like.

Scarlett Johansson confusion

The final unexpected development that hit OpenAI was Scarlett Johansson ScandalSam Altman & Co. approached the actress to lend her voice to GPT-4o. shea film that feels more real than ever. Johansson said no.

OpenAI provided Johansson with an audio of GPT-4o that closely resembled hers, and Altman reportedly tried to get Johansson to reconsider before the big demo, but she refused.

OpenAI released a voice for Skye that looked similar to her, and Johansson contacted her lawyer and made a public statement, with OpenAI retracting the voice of Skye, saying they had used a different actor, but the damage was already done.

First, this should never be done to anyone, especially an actress who plays a beloved Marvel hero who has sued Disney in the past and won, especially in a very public battle that preceded the settlement.

Secondly, given the criticism they have already received for infringing copyright with ChatGPT and other OpenAI products, they are not stealing anything to improve their AI.

OpenAI doesn’t want to become the next Google

Google has done a lot right with its products recently, but it has also done a lot wrong for the sake of profit. It is a huge company that has dealt with a number of scandals over the years, especially regarding privacy violations. These actions have led me to lose trust in Google over the years and to largely remove the company’s flagship products from my computing experience.

This is also why I prefer ChatGPT over Gemini, even though Gemini has many strengths that OpenAI struggles to match.

However, I do not appreciate at all the direction OpenAI is taking. If the pursuit of profit at all costs at the expense of AI safety and common sense is OpenAI’s main priority, I will reconsider my relationship with ChatGPT. I can always switch to another provider, and there are plenty of alternatives.

That being said, OpenAI has time to right the ship before inventing a motto like “don’t be mean” to improve its image.