Not long ago, artificial intelligence was simply a “plus” in our lives. We used it to help with some homework, to find quick answers, or as a translator for documents that we couldn’t understand even if we wanted to.
Now instead, It is no longer just occasional helpthere are now many who admit they cannot live without AI. They use it to look for cooking recipes, psychologist recipes, to prepare for a job interview. and even to know what to reply to WhatsApp messages.
Bill Gates has once again put the focus on the effects that this technology can have on the future of human beings. Gates has identified four specific risks that, in his opinion, You have to pay close attention because they could alter the labor market.
The fear that AI will take our jobs
Automation It is one of the main concerns associated with the advancement of artificial intelligence. More and more administrative tasks can be solved in seconds with the help of algorithmssomething that opens up opportunities, but also generates job uncertainty.
Gates believes that the change will be progressive, although not without difficulties.
In his opinion, some sectors will continue to require human intervention for years, especially those linked to care, education or health. However, he warns that many workers will have to adapt to new functions and acquire new skills so as not to be left out of the market.
Fake videos can have real political impact
Another of the points that most worries the founder of Microsoft is the growth of content manipulated through AI. The so-called deepfakes make it possible to generate increasingly realistic fake videos, audios and images.
On platforms like
A false recording released at the right time, even if it is later denied, can change people’s decisions and raise doubts that are already difficult to correct in time.
AI can also facilitate cyber attacks
Gates warns that AI makes it possible to generate more effective codes, perfect digital fraud and design much more credible deception messagesespecially in type attacks phishing (technique that uses scams, such as emails or calls, to impersonate trusted companies or people)
Access to public data and the speed of these tools increase the ability to personalize these types of traps, something that requires cybersecurity to be further strengthened.
When you respond with confidence… even if you make a mistake
One of the most well-known flaws of AI is its ability to invent convincing-looking data. It’s what happens with so-called “hallucinations”, when systems like ChatGPT fill in information gaps with invented answers that sound coherent.
This happens when he recommends non-existent places, mixes false facts with real ones, or answers with absolute certainty questions whose answer he actually doesn’t know.
To reduce this risk, many experts recommend always asking that information be accompanied by verifiable sources. A useful example is Add specific instructions such as: “If there is no primary source, write ‘unconfirmed data’ and indicate how it could be verified.”
In addition, Gates remembers that these systems can also reproduce biases already present in the data with which they learn, amplifying social prejudices related to gender, origin or identity.