There is a lot of demand for artificial-intelligence engineers, but there seems to be a disparity in salaries between countries
US Software Developers Specialising in Artificial Intelligence (AI) The latest global survey by Stack Overflow shows that machine learning (ML) engineers in the US earn significantly more than their counterparts in other countries. The survey of 49,500 responses from 177 different countries found that the median salary in the US for an AI and ML Engineer is $189.500 per year, which is over $50,000 higher than the global average.
According to the survey, UK AI and ML developers earn $149.756, which is almost $40,000 less than their US counterparts. In France, there is a larger salary gap, with a median annual salary of $104,413. In India, AI/ML engineers can earn as little as $17,434 a yearly.
While the survey found that 47% of developers use AI tools on a daily basis, Stack Overflow reported more developers actively distrust AI tools (46%) rather than trust them (33%).
The study found that experienced developers were the most cautious. They had the lowest “highly-trust” rate (2.6%) and the highest rate of “highly-distrust” (20%). This indicates a widespread need for humans to verify for those with accountability roles.
Of the 11,202 developers that responded to questions regarding how they use AI, the majority (54%) said the primary use was to search for answers. Just over a quarter (35%) of developers use AI to find answers. It is used to create content or synthetic data by 33% of users, and to learn new concepts by 33%.
Another main use is documentation. 31% say they use AI to documentation their code, and a quarter use AI to create or update documentation. Around a fifth use AI to debug and learn about the codebase.
Based on the responses from 31,476 developers, “AI solutions almost right but not quite” is the most common frustration (66%) among those polled. This often leads to the second biggest frustration, “Debugging AI generated code is more difficult,” at 45%.
In the future, when AI is advanced, three quarters of developers say they will still ask for human help if they don’t trust the answers given by AI. According to Stack Overflow this Positions human developers to be the ultimate arbiters for quality and correctness.
The growing lack of confidence in AI tools was the key data point for us in this year’s study, especially with the increased adoption and growth of these AI Tools,” said Prashanth Chandrasekar, CEO of StackOverflow. “AI is an extremely powerful tool, but there are significant risks associated with it, such as misinformation or a lack of complexity or relevance.
With AI becoming more and more prevalent, and ‘AI Slop’ replacing content online, it is important to take a responsible approach that relies on the use of data from curated databases. We believe that by providing a trusted layer of human intelligence in the age AI, the tech enthusiasts of today will be able to play a greater role in adding value in order to build the AI products and technologies of tomorrow.
by Andy Patrizio.
by Andy Patrizio.
