Many companies are implemeing various artificial ielligence tools in their activities. But a new study by MIT shows that most of these experimeal programs fail.
5 % of these programs do not achieve their targeted performance, not because artificial ielligence models do not work as expected, but because general artificial ielligence tools such as ChatGPT are not compatible with workflows in corporate environmes.
Most companies do not properly use artificial ielligence tools
The findings of this study showed that only about 2 % of artificial ielligence test programs lead to rapid acceleration in revenue growth. The report said that most of these projects are halted and have a small or even unnecessary impact on profits and losses. These results are based on 2 ierviews, one survey of 5 employees, and analysis of 2 public artificial ielligence implemeation in companies.
The remaining 2 % is differe from the rest, because they focus on one topic and execute it well. “Some of the test programs of large companies as well as young startups are really shining in the field of productive artificial ielligence,” said Aditia Chalapali, a MIT researcher and author. “The reason is that they choose a specific pain poi, run it well, and ielligely form partnerships with companies that use their tools.”
Another problem that many artificial ielligence organizations face is to determine the wrong priorities for using these tools. The study showed that artificial ielligence is the most efficie in automation of support, which is the same as the administrative and repetitive tasks that many companies usually outsource. However, more than half of the costs spe on artificial ielligence projects focus on sales and marketing; Parts that reasonably require human presence, especially where most buyers are still human, not cars.
The study also found that many companies have not hired new people for job opportunities such as customer support and administrative jobs that have recely been empty and are seeking to use artificial ielligence instead of human employees.
