One problem with this approach of seeking knowledge is the theoretical nature of the question itself:
Unprotected fellatio, which has been practiced by all civilizations since mists of time, is now becoming a cause of concern due to the AIDS epidemic. Most of the sexually transmitted infectious diseases are concerned by fellatio and only few medical studies deal with this topic. This paper is therefore a non exhaustive review of risks brought upon by unprotected fellatio. It is almost impossible to assess the exact risk for a given infection because of the complexity of sexual intercourse, which is rarely exclusively oro-genital.
Therefore: if the following quotes contain numbers, please read them as illustrative and explaining, not as an apology or absolution.
Two such articles that nevertheless do put a number on it in the way you are looking for are:
Assessing the risk of HIV infection after an isolated exposure incident:
The higher the plasma viral load of an HIV-infected person, the greater the risk that their blood, sperm and vaginal secretions are infectious. The risk of infection from splashes onto non-intact skin or a mucous membrane is virtually non-existent if the exposed area is washed within 15 minutes. There is almost no risk of HIV transmission from nasal secretions, saliva, urine and vomit, unless they contain visible blood. The risk associated with sexual exposure to HIV varies enormously, depending on the type of sexual activity: almost non-existent for insertive fellatio; estimated at 1.5% for passive (receptive) anal intercourse with ejaculation inside the rectum. The risk of HIV transmission following injury from a sharp object contaminated with blood has been estimated at about 0.3%.
Risk of HIV infection attributable to oral sex among men who have sex with men and in the population of men who have sex with men.
We examined HIV infection and estimated the population-attributable risk percentage (PAR%) for HIV associated fellatio among men who have sex with other men (MSM). Among 239 MSM who practised exclusively fellatio in the past 6 months, 50% had three partners, 98% unprotected; and 28% had an HIV-positive partner; no HIV was detected. PAR%, based on the number of fellatio partners, ranges from 0.10% for one partner to 0.31% for three partners. The risk of HIV attributable to fellatio is extremely low.
But these numbers might be quite misleading. These are statistical statements! They are calculated for the whole population and do not correspond to what a single individual will do or receive (that is "get").
Instead of finding a number to put the mind at ease a more comprehensive risk reduction strategy might be a much better choice:
Reducing the risk of sexual HIV transmission: quantifying the per-act risk for HIV on the basis of choice of partner, sex act, and condom use:
Sexual acquisition of HIV is influenced by choice of partner, sex act, and condom use. However, current risk-reduction strategies focus mainly on condom use.
Goal: To estimate the contribution of choice of partner, sex act, and condom use on the per-act relative and absolute risks for HIV infection.
Study Design: Per-act relative risk for HIV infection was calculated with use of estimates of HIV prevalence, risk of condom failure, HIV test accuracy, and per-act risk of HIV transmission for different sex acts. Absolute risks were calculated on the basis of these relative risk estimates.
Results: Choosing a partner who tested negative instead of an untested partner reduced the relative risk of HIV infection 47-fold; using condoms, 20-fold; and choosing insertive fellatio rather than insertive anal sex, 13-fold. Choosing one risk-reduction behavior substantially reduces absolute risk of HIV infection for heterosexuals but not for men who have sex with men.
Conclusion: Clarifying the magnitude of risk associated with different choices may help people make effective and sustainable changes in behavior.
Even a small risk is still a risk and in case of nuclear power plants these numbers are big enough to demand that these plants be abolished.
To put that into absolute numbers:
Estimating per-act HIV transmission risk: a systematic review
Sexual exposure risks ranged from low for oral sex to 138 infections per 10 000 exposures.
Further if we are talking risk assessment in the field asked in the question, let us not forget that HIV may be the only concern expressed here, but it is by far not the only illness that should be of concern! Some examples of the "let's not forget" category are:
Oral sex and the transmission of viral STIs:
To review the literature on the role of oral sex in the transmission of viral sexually transmitted infections (STIs).
Conclusions: Oral sex is a common sexual practice among both heterosexual and homosexual couples. The evidence suggests that HIV transmission can take place through oro-genital sex from penis to mouth and vagina to mouth. Case reports describe apparent transmission from mouth to penis although this appears less likely. The risk of oro-genital transmission of HIV is substantially less than from vaginal and anal intercourse. Receptive oro-genital sex carries a small risk of human papillomavirus infection and possibly hepatitis C, while insertive oro-genital contact is an important risk factor for acquisition of HSV 1. Oro-anal transmission can occur with hepatitis A and B. The transmission of other viruses may occur but is unproved. The relative importance of oral sex as a route for the transmission of viruses is likely to increase as other, higher risk sexual practices are avoided for fear of acquiring HIV infection.
Please re-read the last line in bold a few times, to get a feeling of how to better interpret the numbers given above.