For Mustafa Suleyman, head of AI at Microsoft, the situation regarding artificial intelligence is quite clear, although few share this very pessimistic view.
It states that if your workday consists of sitting in front of a computer, opening email, managing projects or writing reports, you may soon be left out. This is what he stated in an interview with the Financial Times. According to the manager, humanity is entering a phase where AI will not only help you, but will completely automate entire classes of jobs.
Suleyman has gone even further, advancing even dates. He believes that within just 12 to 18 months, AI will achieve human-level performance in almost any professional task..
As he states, what previously required years of study and experience to be managed by a human brain is now being processed by algorithms that learn at a speed that no master’s degree or university degree can match.
He ensures that in software engineering, a massive part of the work no longer comes from the minds of programmers, but from AI that writes the code while humans only review and supervise the process.
In the interview with Roula Khalaf, editor of the Financial Timesgave names and surnames to the victims of this automation: lawyers, consultants, marketing managers and project managers. According to the CEO of Microsoft AI, these profiles will disappear due to artificial intelligence in the blink of an eye because their tasks, although complex, are based on processing information, data and texts.
A vision of catastrophic AI that collides head-on with reality
But not everyone buys this speech that seems taken from a movie, which even throws out completely unrealistic dates and which puts, once again, humanity in a danger that certainly does not exist.
While Suleyman and other experts such as Sam Altman of OpenAI or Jensen Huang of Nvidia compete to see who can get the most alarmist headlines (partly because they are interested), the reality of companies is very different.
There is a huge gap between what AI can do in a lab and what actually works in a real office. In fact, while this superintelligence is being talked about, companies like OpenAI are burning money non-stop, even being forced to place ads on ChatGPT to try to make the accounts balance.
Furthermore, there are studies that are beginning to puncture this bubble of optimism that has been carried over these years. A recent report from the Harvard Business Review states that the AI is not reducing the workload; is taking it to levels that border on the unsustainable.
The study, which lasted eight months at a technology company, analyzed how employees used tools such as ChatGPT or Copilot in their daily lives. The result was that workers began to go faster, reach for tasks that did not belong to them, and stretch their workdays until the boundaries between the office and personal life disappeared.
In a simple way, if before you needed more time to do your tasks and you took breaks in between, now that moment of freedom becomes being able to do more work.
In the end, the AI manages to put you in a vicious circle. As everything moves faster, employees rely more on AI to keep up. And since they depend more on AI, they launch into projects for which they are not prepared.
Despite these bumps, Microsoft and OpenAI are moving forward with pharaonic projects like Stargate, a $100 billion supercomputing infrastructure designed specifically to achieve that dream AGI.
Suleyman, for his part, is clear that the future is automatic, but the path to get there seems quite complex to achieve and is not something that can be specified by releasing specific dates. For many there is no doubt that these catastrophic messages are, ultimately, a way to inflate the value of their own companies.