In this regard, Pierluigi Collina, chairman of the FIFA Referee Committee, also responded:
“The test was a major success and we are very confident. In Qatar we will have a very valuable aid to help referees and assistant referees make the best and most correct decisions on the field of play. I know there will be people who say It’s a ‘robot offside referee’: but it’s not. Referees and assistant referees are still responsible for decisions on the playing field.
Controversy over “offside of body hair”
According to the official introduction, SAOT is a system based on artificial intelligence cameras, and 10 to 12 more dedicated cameras will be installed in stadiums using this technology. The transmission speed returns the data, hands it to the software for real-time processing, and sends the processed information to the VAR referee, who then makes the final judgment based on the data.
Semi-automatic offside technology greatly shortens the judgment time. Globally, an offside check for a VAR takes an average of 70 seconds. With the semi-automatic offside technique, the referee only needs to obtain the player’s limb position by visually checking the automatically selected tee point and the automatically created offside line to verify the proposed decision. This process happens in seconds, which means offside decisions can be made faster and more accurately. However, complex events or multiple events occurring at the same time may require a more thorough review of the situation.
This semi-automatic VAR judgment technology based on artificial intelligence bone tracking technology has been used in green fields as early as 2016. With the development in recent years, it has achieved good results: the number of misjudgments has been greatly reduced. But this technology has brought a new controversy to the offside rule – “body hair offside”.
“Body hair offside” refers to those subtle offsides that are difficult to identify with the naked eye but are recognized by VAR. Such offside is very easy to hit the mentality of players and fans. Imagine the players and fans who have been immersed in the joy of scoring goals, but finally poured cold water on this kind of offside decision that is difficult to identify with the naked eye.
Too high recognition accuracy is sometimes a “sin”.
AI has now surpassed human capabilities in many fields, such as image recognition, DNA synthesis prediction, etc. Especially in recent years, with the heavy investment of AI giants, with the support of large model neural networks, writing, programming and other innate human abilities that were originally considered difficult to be surpassed by machines have also begun to be “encroached and captured”. dangerous situation.
To say that the first time the Chinese people have seen the horror of AI, I am afraid it is the battle of the century in which AlphaGo defeated mankind in 2016. The national Go player Ke Jie’s three AI battles were all beaten by AI to a bleak end. In a later interview, when talking about artificial intelligence, Ke Jie said that artificial intelligence is cold, so it is very painful: “Playing chess with AI, the line of winning rate is like a stock market falling limit, and then a horizontal line… “He also said: “In the future, there may be human beings who can defeat it, but it is impossible for me in this life.”
Of course, just as humans can never compare flight with birds, in today’s technologically advanced world, comparing brain power with computing power is definitely not a competition in one latitude, and it is an unfair thing in itself.
But what I want to say here is that it is the “superhuman” in some aspects that allows AI to have many uses.
Machines are created by humans and destined to be born for human needs. Therefore, AI is destined to be only a high-level tool for human beings. In production and life, its positioning is a “good helper”, rather than a dominant player, replacing humans to conquer the world.
That’s why AI can’t become a C-rank even if it wins the rank. In fact, we are not worried about “being replaced”, but people need to define the degree of AI’s involvement in production and life.
Just like in this World Cup, the introduction of VAR was intended to make the referee’s decision more fair and equitable, but can this really ensure this? This is what the fans question the most.
Netizen gadabaj believes: Even so, this will still be controversial. I also like to watch ice hockey, it’s so fast, I can’t see where the puck is most of the time. I’d be happy if the AI could figure out the location and keep showing the puck on the TV screen.
In response, the recent FIFA Football Live broadcast also showcased the semi-automatic offside technology that will be used at the upcoming 2022 FIFA World Cup™ in Qatar. Pierluigi Collina, Chairman of the FIFA Referee Committee, and Johannes Holzmüller, Director of Football Technology and Innovation, explain the adaptability of the technology: If video match officials disagree with the kick/offside line proposed by the system, they can manually select the kick point and use the current There are tools to draw offside lines.
The intervention of AI
The voices of people questioning the AI referee have always existed. Based on various subjective and objective factors, under the condition of only human referees, it is more difficult to make accurate judgments. AI referees can better handle some common disputes, such as whether to score, whether to go out of bounds, whether to be offside, etc. .
In terms of technical maturity, although the current AI football referee is still in the early stage of development, the research and development of core technologies is basically complete. Typical technologies that can be used for AI football referees include but are not limited to: image segmentation processing based on YOLACT, which can be used to automatically determine whether the ball is in the net or out of bounds; based on OpenPose human gesture recognition to determine whether a player is offside or fouled; based on YOLO goals Detection, to achieve ball/player tracking; based on C3D action recognition, to provide accurate basis for handball and other penalties…
When it comes to VAR technology, it was criticized in its early promotion. This kind of AI intervention will interrupt the smoothness of the game, and even make it less enjoyable to watch because it is “too accurate”. “The generally accepted principle is ‘minimum interference, maximum benefit’, that is, ‘minimum interruption of the game, maximum benefit’. As far as the original intention of AI referee research and development is concerned, it is mainly to provide referees with reference information and assist in making decisions. rather than completely replacing the referee.”
What’s more, AI is not omnipotent. As Ren Bin said, “Many referees are also important participants in the project themselves, and their function is not just scoring. For example, a good football referee can control the rhythm and drive the atmosphere, and sports like figure skating involve Aesthetics, whose judgment is subjective, is unlikely to be completely replaced by AI.”
On the one hand, AI is ultimately born for human needs. On the other hand, in practical applications, while AI brings convenience, it will also cause new problems that require new rules to adjust and correct.
Since the birth of the concept of AI, more than half a century of exploration has proved that machines can never truly be like people. Therefore, even in recent years, such as “DeepMind: AI defeated human beings and designed a better economic mechanism”, “AlphaFold2 appeared in Nature again, practitioners are stunned: 98.5% of human proteins have been predicted again” and so on unilaterally surpassed The news of human beings is exposed from time to time, and we will not be jealous and fearful.
On the contrary, “I’m not worried that machines are becoming more and more like people, I’m worried that people are becoming more and more like machines.” Apple CEO Cook said. This kind of worry is necessary. The ubiquitous AI will make people more and more dependent on the judgment of the machine, so that the original human temperature will gradually cool down, and even become icy cold.
Functionalists firmly believe that AI consciousness will eventually awaken, while realists have been trying to remind AI researchers that they need to think about how to reduce the “side effects” brought about by AI in practice, as well as the accompanying inconveniences to AI. sense of trust. This is often an issue that is often overlooked when implementing this high-tech technology.
GIPHY App Key not set. Please check settings