AI colleagues will be on duty! Experts: 15% of daily work will be autonomously completed by AI agents by 2028

The development speed of artificial intelligence is continuously accelerating, and the abilities that were unimaginable in the past have now become a reality. Especially AI agents - or virtual colleagues, in the future, they will work with us and even be able to independently perform tasks.

In fact, Gartner predicts that by 2028, at least 15% of daily work decisions will be autonomously made by artificial intelligence agents (this percentage is still 0% in 2024).

To further emphasize the potential of AI technology, the company has listed it as a top strategic technology trend for 2025. Gene Alvarez, distinguished vice president analyst at Gartner, said, 'This happened very, very quickly.'

No one can finish all the work before going to bed at night, and companies also need to spend a lot of time monitoring things. Creating artificial intelligence agents can not only replace some work, assist companies in detection, but also improve work efficiency and save time.

What other predictions does Gartner have for future strategic technologies? Here are some trends the company discussed at the Gartner IT Symposium/Xpo2024 conference.

AI agents evoke mixed emotions

Alvarez explained, "The primary function of artificial intelligence agents is to solve tedious tasks that consume human time and energy. The next level is the ability for AI agents to autonomously monitor and manage systems."

Image source: Foresight News

"Artificial intelligence agents have the ability to plan, perceive and take action, analyze, repair and report events, rather than just exist as surveillance systems."

In more complex scenarios, agents can one day help improve employee skills. For example, situations that originally required new employees to learn from human colleagues can now be guided by AI colleagues.

Alvarez admits that all of this is both exciting and worrisome, and people's fear of unemployment still exists.

But if an AI agent can truly teach me a new skill, I can transition from a job that is about to be lost to one that is needed.

Establish trust in artificial intelligence

Alvarez then pointed out the next major trend: 'Now there is a whole new workforce, how do we manage them?'

Based on management topics, a series of artificial intelligence governance platforms will be born, enabling companies to manage the legal, ethical, and operational performance of their artificial intelligence systems.

The new tool will create, manage, and enforce policies to ensure transparency in artificial intelligence. These platforms can also check whether AI assistants will have biases and provide information to build models.

Alvarez expects these tools to eventually become part of the artificial intelligence creation process to ensure that ethics and governance are embedded in the model from the outset.

Image source: Foresight News

Alvarez said, "We can create trust through transparency. If people lose trust in artificial intelligence, they will not use it again."

Not just a computational model

Alvarez pointed out that there are seven computing modes currently being used, namely: CPU, GPU, Edge Computing, Application-Specific Integrated Circuit, Neural Morphic System, Classical Quantum Computing, and Optical Computing.

Alvarez said, "We have always had an attitude of constantly adapting to change, but the future hybrid computing model will combine different computing, storage, and networking mechanisms to operate. Coordination software will transfer computing from one to another according to tasks and the most suitable method for the task."

"At the same time, new and more special computing technologies will greatly reduce energy consumption. This is very important because the pressure of energy consumption and carbon footprint is increasing. But with it, the demand for IT computing power by artificial intelligence is also growing at an astonishing rate."

"Incremental improvements are not enough, enterprises need long-term solutions. They also need new technologies, such as green cloud service providers or more efficient algorithms that could potentially increase efficiency by several k times, or even tens of thousands or even hundreds of thousands of times."

Actively address information security issues

Artificial intelligence allows some 'threat actors' to spread false information faster and easier. They can push fabricated information on the Depth, create phishing emails, exploit vulnerabilities in collaboration tools, use malware for eavesdropping, launch account attacks, etc.

Image source: Foresight News

Therefore, it is crucial to address these false information and protect information security, actively assess the authenticity, track the spread of harmful information, and prevent the emergence of counterfeit information.

Its elements include brand impersonation scanning, third-party content assessment, claims and identity verification, phishing mitigation, account takeover prevention, social/media monitoring, and dark web monitoring, as well as emotional manipulation.

Alvarez explained that Depth forgery detection will also be able to identify synthesized media, and watermarking tools will help ensure user interaction with real people.

According to Gartner's prediction, by 2028, half of the enterprises will start to adopt products, services, or functions specifically designed for false information security, and currently this proportion is less than 5%.

Alvarez said, "False information security will not be just a single technology, it will be a series of technologies."

Prepare for the 'post-quantum era' securely

Currently, the internet uses Public Key encryption technology, namely asymmetric encryption technology, which can ensure the security of communication between two points.

Alvarez explained, 'This encryption is very difficult to crack because the decryption time is too long. However, quantum technology is developing rapidly, and it will come into play at some point because it has the mathematical ability to decrypt instantly.'

Gartner predicts that by 2029, the progress of Quantum Computing will render most traditional asymmetric encryption techniques insecure.

Alvarez added: "Organizations now need to prepare for post-quantum encryption technology to ensure that their data can withstand decryption. Transitioning encryption methods is not easy, and this is 'not a simple patch'."

Image source: Foresight News

The established standards of the National Institute of Standards and Technology (NIST) are a good starting point. Alvarez pointed out that the institution will release the second edition of the quantum encryption guidelines in spring 2025.

When all the locks are broken, what do you do? You need new locks. We need to ensure our security is updated before quantum technology becomes a reality.

Artificial intelligence enhances our cognitive abilities

Gartner is turning its attention to the science fiction industry, predicting that the use of bidirectional brain-machine interfaces (BBMI) will increase. This interface can read and decode brain activity to enhance human cognitive abilities.

Alvarez explained that these devices can be directly integrated into our brains, or achieved through wearable devices such as glasses or headbands.

Gartner predicts that by 2030, 30% of knowledge workers will use technologies such as BBMI to remain competitive in an AI-driven work environment (compared to less than 1% in 2024).

Alvarez said, "I see the potential for human skills improvement and next-generation marketing, for example, brands will be able to understand consumers' thoughts and feelings, thus judging consumers' emotions."

Alvarez ultimately compared it with the 2011 film "Limitless" and the AppleTV web series "Life Cutting Art".

Alvarez said, "To be fair, neither of these works portrays this technology in a positive light, but it enhances people's cognitive abilities by accessing the human brain."

[Disclaimer] The market is risky, and investment requires caution. This article does not constitute investment advice, and users should consider whether any opinions, viewpoints, or conclusions in this article are applicable to their specific circumstances. Investing based on this is at your own risk.

This article is authorized to be reproduced from: 'Foresight News'

Original author: VentureBeat