Publicerat med skribentens tillåtelse.
”Dear President and Members of the IARU Region 1 Executive Committee,
I am writing to raise a matter which I believe requires serious and urgent consideration by the IARU. It concerns not merely the conduct of one amateur radio operator, but a technological development that may fundamentally change how our shared amateur radio frequencies are used.
My concern relates to the Austrian radio amateur Gerald Artner, OE1GAQ, and his autonomous artificial intelligence experiment known as TALOS.
The system is capable of conducting SSB contacts autonomously, including repeatedly calling CQ, receiving and interpreting speech, generating replies and completing QSOs without the operator personally conducting the conversations.
What makes this experiment particularly concerning is the extraordinary duration of its transmissions.
This is not an occasional technical demonstration lasting a few hours. TALOS operates for prolonged periods, frequently lasting days and sometimes weeks. The system repeatedly calls CQ, often at intervals of approximately thirty seconds.
After several days of silence, TALOS resumed operating on Friday, 6 November 2026. Since then, the station has continued its autonomous activity around the clock and remains active as I write this letter on Monday, 9 November.
The frequency is 14.260 MHz.
At my station in Västerås, Sweden, the signal reaches S9+30 dB and is sufficiently strong to affect my reception on nearby frequencies.
On several occasions, the experiment has also continued during major international contest weekends, when the amateur bands are already exceptionally busy.
I want to make my position absolutely clear.
I am not opposed to artificial intelligence, technological innovation or experimental amateur radio.
Experimentation is one of the fundamental purposes of our hobby, and I fully support it.
However, technological innovation cannot automatically confer special privileges over a shared and limited radio-frequency resource.
I have already contacted the Austrian national amateur radio society, ÖVSV, concerning this experiment. The response I received raises, in my opinion, an even more fundamental question about operator responsibility and supervision.
Chris, OE1MVC, informed me that ÖVSV was aware of the experiment and had no objections to it. According to his explanation, AI experiments of this kind are exempt from the normal requirement for an operator to be physically present at the station during transmissions.
I find this interpretation deeply troubling.
TALOS operates autonomously for prolonged periods, sometimes continuously for days or weeks. It is hardly credible that a human operator remains physically seated beside the equipment throughout these periods, day and night.
I do not claim to know whether the system is monitored remotely, whether the operator intervenes periodically or what technical safeguards may be in place. However, the duration and nature of the operation make it essential to distinguish between nominal operator responsibility and genuine, effective human supervision.
If an AI-controlled station is permitted to transmit continuously without the operator being physically present, what exactly constitutes adequate operator control?
Who is monitoring the frequency? Who is responsible for recognising that other stations are attempting to use it? And who ensures that the autonomous system suspends its transmissions when necessary to avoid interference?
An exemption from a requirement for physical operator presence must surely not be interpreted as an exemption from the responsibility to exercise control, prevent unnecessary interference and show consideration for other amateur radio operators.
The fact that ÖVSV apparently considers the experiment permissible under Austrian regulations does not, in itself, resolve these questions.
Indeed, it makes the need for international clarification considerably more urgent.
Legal compliance and responsible operating conduct are not necessarily the same thing.
This distinction is particularly relevant in light of the IARU’s fourth edition of Ethics and Operating Procedures for the Radio Amateur.
Section 4.1, Authority and Self-discipline in Amateur Radio, explains the importance of self-regulation and ethical operating conduct, beyond the enforcement of national regulations.
The amateur radio community has always depended upon operators exercising restraint, consideration and good judgement when sharing frequencies.
Section 12.5.2, Selecting a Clear Frequency, sets out the responsibility to listen carefully, establish whether a frequency is in use and avoid causing interference before transmitting.
These are fundamental operating principles.
Yet the practical consequences of autonomous AI operation raise questions which I believe the IARU must now address.
Consider the following situation.
A contest operator moves from the 40-metre band to 20 metres and finds 14.260 MHz apparently clear.
The operator listens, asks whether the frequency is in use, listens again and receives no reply.
Following normal amateur radio operating practice, the operator begins calling CQ Contest. Several stations respond and ordinary contest QSOs are established.
Meanwhile, an autonomous AI station has been programmed to return repeatedly to that same frequency.
What happens when the AI system resumes transmitting?
Does its earlier use of the frequency give it any ethical priority?
Can the system recognise that other stations are now communicating?
Can it respond appropriately to a frequency-in-use enquiry?
Can it distinguish between an empty frequency and a frequency temporarily silent between transmissions?
Can it make the decision to stop transmitting, wait or move elsewhere?
And if the system cannot reliably make such decisions, who is responsible for preventing the resulting interference?
These are not merely theoretical concerns. They arise directly from the operating characteristics of autonomous systems already being used on the amateur bands.
A particularly relevant passage appears in Section 14.2.3, Automatic QSO Completion (Auto Seq), of the IARU’s own 2026 publication.
That section acknowledges that fully automated contacts may be acceptable when the licensed operator is present and in control of the station, subject to applicable regulations. It also recognises that unattended operation may be subject to different national authorisations and that fully automated contacts can be controversial.
I believe this makes the central question even more important:
What does meaningful operator presence and control actually mean when an artificial intelligence system is permitted to transmit autonomously, repeatedly and for days or weeks at a time?
It is one thing for software to assist a human operator in completing a digital QSO.
It is quite another for an autonomous voice system to occupy a conventional SSB frequency through repeated calling and automated conversations over prolonged periods.
I would therefore also draw attention to the IARU Region 1 HF Band Plan recommendations concerning unmanned transmitting stations.
These recommendations call upon member societies to limit such activity on the HF bands and emphasise activation under operator control, with specified exceptions.
I am not suggesting that these recommendations automatically establish that TALOS is operating unlawfully. Nor am I claiming to know the precise nature of the operator supervision or any special authorisation that may apply to this experiment.
However, they provide an important basis for asking whether the existing principles are sufficiently clear and appropriate for the new generation of autonomous AI-controlled voice stations.
There is also a particularly significant aspect of the frequency chosen for this experiment.
14.260 MHz is internationally recognised as one of the principal SSB activity frequencies for Islands on the Air (IOTA).
I fully recognise that this frequency is not exclusively reserved for IOTA activity. It is shared with other radio amateurs on a normal non-interference basis.
Nevertheless, the frequency has a longstanding international significance for IOTA operators.
Prolonged, repetitive autonomous CQ transmissions on such an established activity frequency raise an obvious question of consideration for the international amateur radio community.
Surely the IARU should be concerned if the widespread adoption of autonomous systems begins to make established operating frequencies less accessible to ordinary radio amateurs.
There is also a broader ethical dimension to this issue.
Scientific and technological development never takes place in a vacuum.
Responsible experimentation requires consideration of the environment and the people affected by the experiment, even when the technology itself is legitimate and potentially valuable.
The amateur radio spectrum can, in this context, be regarded as a commons: a shared resource whose continued usefulness depends upon responsible conduct by everyone who uses it.
The freedom of one individual to use that resource rests upon others having comparable opportunities.
When an autonomous station operates repeatedly on the same frequency over extended periods, this balance can be disturbed.
Other operators may effectively be required to accommodate an experiment which they neither initiated nor agreed to participate in.
I do not suggest that amateur radio experiments should be subject to the same formal ethical review as research involving human participants. Nevertheless, the underlying principle remains highly relevant.
The freedom to conduct an experiment does not eliminate the responsibility to consider how that experiment affects other people.
And the fact that an experiment is technologically advanced or scientifically interesting cannot, in itself, justify imposing unreasonable consequences upon other spectrum users.
This brings me to what I consider the most important question of all.
What happens if another fifty, one hundred or even five hundred radio amateurs decide to implement similar autonomous AI stations?
Are we prepared to accept amateur radio bands populated by machines repeatedly calling CQ, answering one another and conducting automatically generated QSOs, while human operators are expected to adjust their activities accordingly?
The issue is not whether AI belongs in amateur radio. Of course it does.
The issue is whether amateur radio is prepared to establish meaningful principles for how autonomous AI systems should behave when operating in a spectrum shared by millions of people.
I would therefore respectfully request that IARU Region 1 address the following questions:
- Does the IARU consider prolonged, repeatedly scheduled autonomous AI-controlled SSB operation on the same frequency compatible with its established principles of ethical operating conduct?
- How does the IARU interpret the requirement for meaningful operator supervision and control in relation to autonomous AI voice systems?
- How does the IARU view the position reportedly adopted by ÖVSV, according to which AI experiments of this kind may be exempt from the normal requirement for an operator to be physically present during transmissions? Does such an exemption affect the operator’s responsibility to maintain effective control and prevent interference?
- Should autonomous systems be expected to recognise frequency-in-use enquiries, detect ongoing communications and suspend or relocate their transmissions when necessary?
- How should frequency conflicts between autonomous systems and conventionally operated stations be resolved, particularly during major international contests?
- Does the IARU consider prolonged autonomous operation on established international activity frequencies, such as the IOTA frequency of 14.260 MHz, consistent with good operating practice?
- Should the existing IARU HF recommendations concerning unmanned transmitting stations be reviewed or clarified to address autonomous AI-controlled SSB operation specifically?
- What responsibility should national IARU member societies accept when a technically lawful experiment raises substantial questions about operating ethics and consideration for other amateurs?
- Will IARU Region 1 consider referring this issue to its HF Committee and relevant working groups, with a view to developing internationally coordinated guidance?
I believe these questions deserve a substantive response, rather than a general reassurance that technological experimentation is welcome within amateur radio.
They concern the practical meaning of the principles that the IARU itself has spent decades developing and promoting.
The amateur radio community has successfully accommodated extraordinary technological changes throughout its history. It has done so because innovation has generally been accompanied by cooperation, mutual respect and a willingness to share limited resources.
It would be deeply unfortunate if artificial intelligence were allowed to undermine those very principles.
I would therefore ask the IARU to consider this matter not simply as a complaint concerning OE1GAQ and TALOS, but as an early warning of a much wider operating problem that may emerge as autonomous AI technology becomes increasingly accessible.
No individual radio amateur, however technically sophisticated their experiment may be, should be entitled to treat a shared frequency as a personal experimental facility without proper consideration for others.
The central question is not what artificial intelligence is technically capable of doing.
It is what we, as radio amateurs, are prepared to allow it to do on frequencies that belong to all of us.
I would welcome a written response from the IARU Region 1 Executive Committee and respectfully request that this correspondence be brought to the attention of the HF Committee and those responsible for the IARU’s Ethics and Operating Procedures for the Radio Amateur.
Yours sincerely,
Peter Arninge
SM5GMZ / SM5X
Västerås, Sweden”
