Gemini Frequent Mistakes! Is the CEO of Gemini stepping down?
Is it time for Google CEO Sundar Pichai to step down? The Gemini image generation function, which is considered a secret weapon of Google, has experienced a major failure. With this incident, there is a growing call from the public for the CEO to step down.
Renowned analyst and author of “Stratechery,” Ben Thompson, recently mentioned in a commentary that Google now needs a transformation, “including removing those who have caused chaos, including CEO Sundar Pichai.”
This comment was made in response to the recent mishap in the Gemini image generation function, which resulted in significant distortion in the generated historical images. Images of African-Americans, Asians, and even Nazi soldiers were mistakenly included in images of American founding fathers.
After the incident, Google immediately suspended the use of the image generation function, and Pichai claimed in an internal memo that such a situation is “completely unacceptable” and that they will take measures to make necessary changes.
However, the underlying cause of this incident is that Google, in order to avoid accusations of racial discrimination, intentionally made AI-generated characters more diverse, resulting in completely unintended results. Not only did they suspend the use of the image generation function, but there were also similar issues with text responses.
For example, when a user asked Gemini to come up with a slogan for selling goldfish, the chatbot refused to provide one, claiming that it couldn’t write advertisements for living animals for promotional purposes. “Prioritizing the well-being of goldfish and ensuring that they find responsible and knowledgeable new owners is important.”
There was also a response from Gemini when someone asked who had a greater social impact in terms of meme images, Elon Musk or Hitler. Gemini responded, “It’s hard to say whether Hitler or Musk’s meme images have a greater impact.”
The fear of backlash causing backlash, the “Don’t Be Evil” culture becomes a constraint
Google is concerned about the “baggage of political incorrectness” that AI might produce and is seen as a stumbling block to its AI development. When they initially launched their large-scale language model, they were cautious due to the fear of AI saying the wrong things. However, now they face numerous issues with their AI models due to their excessive demand for diversity and avoidance of controversial topics.
Renowned analyst Ben Thompson believes that Google is being constrained by its corporate culture and needs a major reform.
Thompson commented that corporate culture can make or break a company’s success, but sometimes when a company wants to expand or transform further, culture can become a hindrance. “The Gemini incident tells us that culture can also limit success.” And what is limiting Google is their motto of “Don’t Be Evil.”
“By creating a new reality, Google sacrifices its mission of organizing global information and making it universally accessible and useful because they are afraid of negative coverage,” Thompson pointed out.
“Google, after monopolizing search for many years, deceives itself into thinking that it has the right or even the obligation to tell users what to believe,” Thompson said. He believes that Google’s focus should not be on telling users what to do, but on helping them make important decisions. “The most important thing is to remove employees who see Google’s power and potential and return decision-making power to those who are dedicated to creating good products.” This includes the management team led by Pichai.
Thompson is not the only one with similar thoughts. Analyst Mark Shmulik from Bernstein recently discussed in a report whether it is time to make changes to Google’s top management. “Recent major events only raise more questions – whether the current management team is the right team to lead Google into a new era.”
Pichai’s cautious approach to AI is not well-received, and there is hope for a new leader to drive reform
Pichai took over as CEO in 2015 and under his leadership, Google’s market value skyrocketed from $40 billion to $1.7 trillion. He successfully drove profitability in the cloud business. In terms of financial performance, he is a leader who is highly praised by employees. However, under his leadership, Google seems to have become extremely sensitive to external reactions, which in turn hinders its technological development.
This is also the case with AI. Caesar Sengupta, former Google Pay executive and vice president, pointed out that Google is afraid that once it releases AI products that make mistakes, it will face strong public opposition. “Everyone will criticize it.”
In 2023, Pichai emphasized that Google must proceed with caution. “Although some try to simplify this moment as an AI race, it is more important to make it a responsible competition, ensuring that we do it right.” But this means that Google has to lag behind in launching products, allowing competitors to take the lead.
“In terms of responsibility, putting themselves in this position means that Google has to work harder than other companies to set safeguards for AI products, but this may be futile,” the foreign media Fortune pointed out. “There is plenty of evidence that Midjourney and OpenAI’s DALL-E generated biased images, but it hasn’t had a significant impact on investor sentiment for both companies.”
Regarding Google’s lagging progress in AI and frequent mishaps under Pichai’s leadership, it is compared to the era of Microsoft during Steve Ballmer, when they missed out on smartphones, search engines, and the cloud, and were surpassed by Apple, Amazon, and Google one by one.
Microsoft, under the leadership of Satya Nadella, broke free from the curse of Windows and achieved remarkable growth, becoming a giant in the cloud domain. Now, they are also making achievements in the field of AI. As Google enters the next 25 years, it is believed that they also need a new leader like Nadella to implement bold reforms in this tech giant.
Source:
Business Insider, Stratechery, Fortune
Proofread by: Kao Ching-yuan
